Copy Link
Add to Bookmark
Report
Alife Digest Number 015
Artificial Life Digest, Number 15
Monday, April 9th 1990
Issue's Topics:
re: GAs and binary coding
binary encoding & evolution
Re: Artificial Life Digest, #13
RE: measures of entropy in alife models (ComMet)
please post
Cellsim2.5 FTP change and bugfixes
----------------------------------------------------------------------
Date: Fri, 6 Apr 90 10:40:36 EDT
From: "David M. Chess" <CHESS@ibm.com>
Subject: re: GAs and binary coding
(Of course, unless the crossover and other mutation operators operate
on the bits within the numbers of the n-ary encoding, you don't get
the evoluationary effects. That's perhaps another way of expressing
the value of a binary representation: every bit (in both senses) of
the information in the representation is available for evolution to
operate on. DC )
------------------------------
Date: Fri, 6 Apr 90 10:43:45 -0500
From: Marek Lugowski <marek@iuvax.cs.indiana.edu>
Subject: binary encoding & evolution
I am tempted to question David Chess' assertion on the value of a binary
representation (contained elsewhere in this digest, I hope, or one of
the immediately previous messages on the reflector).
What David describes is an artifact of (1) typography, (2) geometry implicit
in the typography of bit string manipulations, (3) discreteness of
representation as symbolic logic.
I am thinking of candidate media for evolution to tweak every bit (in a
bouquet of senses, not just two...) which are neither binary nor
typographic:
1. energy levels in quantum wells. See literature on the nascent quantum
mechanics-driven solid state devices (e.g., BiQQRT) out of TI, ATT
and various universities.
2. in vivo genetics/in vitro chemical computers. ATT, U Miami &?
3. discrete geometry embodiments. Our (Indiana U.) Computational
Metabolism (ComMet) with its N-sided tiles of M colors and O bits of
state per color (and P distinct geometries on R dimensions of tiling) is
one candidate. We have not gotten around to messing (smile) with GAs
there yet. Any help accepted with gratitude. We'll tell you everything
we know so far about ComMets, which is scant (smile).
-- Marek
marek@iuvax.cs.indiana.edu
------------------------------
Subject: Re: Artificial Life Digest, #13
Date: Fri, 06 Apr 90 09:26:31 -0700
From: "David A. Honig" <honig@bonnie.ics.uci.edu>
Re: whether it is 'bad' to have many genotypes code the same phenotype:
There was an article (summarized in Sci News, from Science I think) about
how some computational chemists found that *many sequences* of amino
acids yield the same shape when the protein folds. Now, no doubt there
may be critical sequences, but their point was that nature gives you
some slack (my interpretation), since shape matters a lot for enzymatic
activity.
So it may be not only useful but perhaps realistic to have redundancy in
geno to pheno expression.
------------------------------
Date: Fri, 6 Apr 90 17:06:49 EDT
From: mclennan%MACLENNAN.CS.UTK.EDU@cs.utk.edu
Subject: RE: measures of entropy in alife models (ComMet)
Marek Lugowski writes:
> I am aware of measures of entropy in cellular automata work.
>
> Question to the list membership: are these measures sufficient to
> characterize the relevant change in disorder in life simulations?
>
> Would you happen to have a candidate measure of entropy that significantly
> differs from those used in cellular automata to share with us?
Be careful to distinguish entropy, which has a very specific meaning, from
other measures of structure or order (or their opposites). Entropy is a
property of a probability distribution $p_k$ , which is defined
$H = - \sum_k p_k \log p_k$
Two of its most significant properties are that it's additive and that
it's maximized by a uniform distribution. Thus, if a uniform distribution
corresponds to "disorder" in your problem, then entropy is a measure of
"disorder." You can apply entropy anywhere you have a probability
distribution in which order corresponds to deviation from uniformity.
For example, you can compute an entropy for each cell by keeping track of
the frequencies with which the colors appear. Entropy decreases as the
color becomes more constant.
For other measures of order/disorder you might look to statistics (e.g.,
chi-squared). These will allow you to quantify how much the present
state of a cell allows you to predict its own future state, or the
future state of neighboring cells.
Hope that's some help.
Bruce MacLennan
Department of Computer Science
107 Ayres Hall
The University of Tennessee
Knoxville, TN 37996-1301
(615)974-0994/5067
maclennan@cs.utk.edu
------------------------------
Date: Sun, 8 Apr 90 17:45:08 CDT
From: shriver@usl.edu (Shriver Bruce D)
Subject: please post
===============================================================
This note is being separately posted on the following bulletin
boards:
connectionists
neuron-digest
neutran
Please recommend other bulletin boards that you think are also
appropriate.
===============================================================
I am interested in learning what experiences people have had using
neural network chips. In an article that Colin Johnson did for PC
AI's January/February 1990 issue, he listed the information given
below about a number of NN chips (I've rearranged it in
alphabetical order by company name). This list is undoubtedly
incomplete (no efforts at universities and industrial research
laboratories are listed, for example) and may have inaccuracies in
it.
Such a list would be more useful if it would contain the name,
address, phone number, FAX number, and electronic mail address of
a contact person at each company would be identified.
Information about the hardware and software support (interface and
coprocessor boards, prototype development kits, simulators,
development software, etc.) is missing.
Additionally, pointers to researchers who are planning to or have
actually been using these or similar chips would be extremely
useful. I am interested in finding out the range of intended
applications.
Could you please send me:
a) updates and corrections to the list
b) company contact information
c) hardware and software support information
d) information about plans to use or experiences with having used
any of these chips (or chips that are not listed)
In a few weeks, if I get a sufficient response, I will resubmit an
enhanced listing of this information to the bulletin boards to
which I originally sent this note. Thanks,
Bruce Shriver (shriver@usl.edu)
=================================================================
Company: Accotech
Chip Name: AK107
Description: an Intel 8051 digital microprocessor with its on-
chip ROM coded for neural networks
Availability: available now
Company: Fujitsu Ltd.
Chip Name: MB4442
Description: one neuron chip capable of 70,000 connections per
second
Availability: available in Japan now
Company: Hitachi Ltd.
Chip Name: none yet
Description: information encoded in pulse trains
Availability: experimental
Company: HNC Inc.
Chip Name: HNC-100X
Description: 100 million connections per second
Availability: Army battlefield computer
Company: HNC
Chip Name: HNC-200X
Description: 2.5 billion connections per second
Availability: Defense Advanced Research Projects Agency (DARPA)
contract
Company: Intel Corp
Chip Name: N64
Description: 2.5 connections per second 64-by-64-by-64 with
10,000 synapses
Availability: available now
Company: Micro Devices
Chip Name: MD1210
Description: fuzzy logic combined with neural networks in its
fuzzy comparator chip
Availability: available now
Company: Motorola Inc.
Chip Name: none yet
Description: "whole brain" chip models senses, reflex, instinct-
the "old brain"
Availability: late in 1990
Company: NASA, Jet Propulsion Laboratory (JPL)
Chip Name: none yet
Description: synapse is charge on capacitors that are refreshed
from RAM
Availability: experimental
Company: NEC Corp.
Chip Name: uPD7281
Description: a data-flow chip set that NEC sells on PC board
with neural software
Availability: available in Japan
Company: Nestor Inc.
Chip Name: NNC
Description: 150 million connections per second, 150,000
connections
Availability: Defense Dept. contract due in 1991
Company: Nippon Telephone and Telegraph (NTT)
Chip Name: none yet
Description: massive array of 65,536 one-bit processors on 1024
chips
Availability: experimental
Company: Science Applications International. Corp.
Chip Name: none yet
Description: information encoded in pulse trains
Availability: Defense Advanced Research Projects Agency (DARPA)
contract
Company: Syntonic Systems Inc.
Chip Name: Dendros-1
Dendros-2
Description: each has 22 synapses, two required by any number can
be used
Availability: available now
------------------------------
Date: Mon, 9 Apr 90 13:58:43 EDT
From: hiebeler@turing.cs.rpi.edu (Dave Hiebeler)
Subject: Cellsim2.5 FTP change and bugfixes
[ I am sending this to the ALife mailing list, because there may be
several new users on that list who missed the original announcement of
Cellsim 2.5, who would be interested in Cellsim. I apologize to those
who will see multiple copies of this message, as it was also sent to
the Cellsim user-list and the cellular automata mailing-list. -D.H. ]
================
Two pieces of news for Cellsim users. The FTP site for obtaining
Cellsim has changed, and some bugs in V2.5 have been fixed.
================
Anonymous FTP is no longer available on the machine that Cellsim was
previously being distributed from, so we have moved the distribution
files to another site. You can now obtain Cellsim (the current
version is V2.5) via anonymous FTP to turing.cs.rpi.edu (128.213.1.1).
Use "anonymous" as username, and your e-mail address as a password.
All Cellsim files are in the directory "pub/cellsim". The current V2.5
tarfile has had the bugfixes applied to it already. If you got Cellsim
before April 8, 1990, you should read below to see how to fix your version.
***NOTE***
Be sure to transfer files in binary mode (type "bin" at the FTP prompt
to set binary mode). Also, please try to restrict your FTP's to the
non-business hours (e.g. after 7pm, before 8am, Eastern time zone).
**********
If you've never heard of Cellsim before, this is briefly what it is: a
SunView-based cellular automata simulator that runs on color and B&W
Sun-3's, Sun-4's, and Sparcstations. It can run 1-D and 2-D CA, on
array-sizes of up to 512x512, using up to 256 states/cell. It can
also attach to a Connection Machine either locally or through the
network, and perform the computations there much more quickly. A copy
of the Cellsim V2.5 announcement, which has some more information including
details on how to obtain the package, is in the FTP directory as
the file "cellsim_2.5.announcement". If you obtain Cellsim for the first
time, please send a message to one of the authors, so you will be added
to the mailing-list to be notified of future releases or bug fixes.
If you already got Cellsim V2.5 when it was first announced, then you
don't need to generate lots of network traffic by getting the entire
distribution again. You only need to get a few files through anonymous FTP:
1. Get the file "2.5fixes.patch.Z".
2. If you use Cellsim with a Connection Machine, get the file
"2.5CMfixes.patch.Z".
2. Get the file "2.5fixes.README", which explains how to apply the patches
to V2.5.
If you can't FTP, then send a message to hiebeler@turing.cs.rpi.edu, I
will probably be able to mail the files to you (I think they're small enough).
The following bugs were reported in V2.5, and have been fixed:
1. When you did "Run/bound" when attached to CM, the current time was not
updated in the control panel after running.
2. When you changed the CMFE host in the CM defaults popup, it wouldn't take
effect until you closed the popup (similarly for CMFE port). Now it takes
effect immediately, so you can leave the CM defaults popup open if you
like.
3. Couldn't turn off CMFB display while running on the CM.
4. When running skip/bounded on CM, it didn't update Sun display at the end,
even if the Sun display was enabled.
5. "B" wasn't defined in the distributed CMnborhood.h, so if you wanted to
write a Paris rule that needed B, it wouldn't compile.
6. When you tried to dump a raster image while running on a Sun-3, it
complained that it was in closeup mode (even when it wasn't!)
7. "Run/bounded" didn't work right in 1-D (it acted like "screenful").
"Run/skip" and "Run/skip-bound" also didn't work right in 1-D
Also, the following two features have been added:
1. Added ability to load/save 2-d images, when in 1-d mode. Previously, only
the bottom ("current") line of the array would be saved or loaded when in
1-D mode. Now, there is a toggle in the "Image" defaults popup which you
can toggle so that load/save will operate on the entire 2-D image.
2. When running on a Sun, you can now use any array size (up to a maximum of
512) that is a multiple of 4. Previously, the only allowed sizes were
64, 128, 256 and 512. Note that if you're attached to the CM, the only
allowed sizes are still 128, 256, and 512.
In a future release, we hope to have truly arbitrary array-sizes working
on the Sun (not just multiples of 4), and perhaps even on the CM although
that may not be practical.
Note that these two new features are *not* mentioned in the V2.5
documentation. They weren't going to be available until the next version,
but a user needed these features quickly, so I put them in early.
Thanks to everyone who has notified us of bugs in Cellsim V2.5. If
you notice a bug that is not listed here, please let us know. We also
appreciate suggestions for new features to incorporate into Cellsim;
many of the new features in V2.5 were suggested by the user community!
Chris Langton Dave Hiebeler
Theoretical Division hiebeler@heretic.lanl.gov
T-13, MS B213 or hiebeler@turing.cs.rpi.edu
Los Alamos National Laboratory
Los Alamos, NM 87545 USA
cgl@lanl.gov
--
Dave Hiebeler hiebeler@turing.cs.rpi.edu
Center for Nonlinear Studies and Theoretical Division, Los Alamos Nat'l Lab
Computer Science Dept, RPI
------------------------------
End of ALife Digest
********************************
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=---=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
= Artificial Life Distribution List =
= =
= All submissions for distribution to: alife@iuvax.cs.indiana.edu =
= All list subscriber additions, deletions, or administrative details to: =
= alife-request@iuvax.cs.indiana.edu =
= All software, tech reports to Alife depository through =
= anonymous ftp at iuvax.cs.indiana.edu in ~ftp/pub/alife =
= =
= List maintainers: Elisabeth Freeman, Eric Freeman, Marek Lugowski =
= Artificial Life Research Group, Indiana University =
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=---=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
End of Alife Digest
********************************
